[Openmp-dev] Patches to support OMPT

Hahnfeld, Jonas Hahnfeld at itc.rwth-aachen.de
Thu Apr 23 06:06:39 PDT 2015


we have just updated the revisions with your feedback.
Any further comments on this?

Jonas Hahnfeld

> -----Original Message-----
> From: openmp-dev-bounces at cs.uiuc.edu [mailto:openmp-dev-
> bounces at cs.uiuc.edu] On Behalf Of Hahnfeld, Jonas
> Sent: Thursday, April 09, 2015 3:25 PM
> To: 'openmp-dev at dcs-maillist2.engr.illinois.edu'
> Cc: John Mellor-Crummey (johnmc at rice.edu); Cramer, Tim
> Subject: [Openmp-dev] Patches to support OMPT
> Hi all!
> After some months of work, we would like to share our patches to add
> support for OMPT[1] to this version of the Intel runtime.
> Implementation was mainly done by John Mellor-Crummey from Rice
> University in a repository at Google Code[2].
> Support for OMPT is disabled by default and therefore has to be enabled
> explicitly:
> With CMake, define "ompt-support": $ cmake -Dompt-support=true ..
> build.pl works nearly the same: $ make build_args="--ompt-support=on"
> With this switch, ompt-blame and ompt-trace will be enabled as well which
> can be disabled separately.
> We've uploaded the patches to LLVM's Phabricator. Some short explanation:
> 0001-kmp_lock-Change-return-type-of-release-functions-fro.patch[3]
> This is a preparation patch to get information about the actual release
> from nested locks.
> 0002-Add-new-files-for-OMPT.patch[4]
> This patch contains the new files for OMPT and the needed changes to the
> build infrastructure.
> 0003-Modify-runtime-to-call-correct-functions-for-OMPT.patch[5]
> These are the actual changes in the runtime to issue OMPT-related
> All of them are surrounded by #if OMPT_SUPPORT and can be disabled
> (which is the default).
> The implementation is tested with the OMPT-TestSuite[6] and conforms with
> the OpenMP's 2nd Technical Report[1].
> Additionally we ran the OpenMP Validation Suite contained in this
> to make sure that nothing broke.
> Please review and commit if found suitable.
> We know that these are rather big patches (for a big feature), but we
> couldn't come up with a better splitting.
> From Aachen,
> Tim Cramer and Jonas Hahnfeld
> [1]: http://openmp.org/mp-documents/ompt-tr2.pdf
> [2]: https://code.google.com/p/ompt-intel-openmp/
> [3]: http://reviews.llvm.org/D8916
> [4]: http://reviews.llvm.org/D8917
> [5]: http://reviews.llvm.org/D8918
> [6]: https://code.google.com/p/ompt-test-suite/
> --
> Jonas Hahnfeld, MATSE-Auszubildender
> IT Center
> Group: High Performance Computing
> Division: Computational Science and Engineering RWTH Aachen University
> Seffenter Weg 23 D 52074  Aachen (Germany) Hahnfeld at itc.rwth-aachen.de
> www.itc.rwth-aachen.de

-------------- next part --------------
A non-text attachment was scrubbed...
Name: smime.p7s
Type: application/pkcs7-signature
Size: 5868 bytes
Desc: not available
URL: <http://lists.llvm.org/pipermail/openmp-dev/attachments/20150423/d029bdc1/attachment.bin>

More information about the Openmp-dev mailing list